Reactive Shiny flexdashboard Dynamic Inputs - r

Question

I am developing a small app that runs on a server. The data is filtered using a session user id. The radio choices should be conditional on the reactive data (only the values of var1 that exist for user "A" in var3 should be available for choice when user "A" is logged in). Any hint? Is observe and updateSelectInput the way to go?
---
title: "Untitled"
runtime: shiny
output:
flexdashboard::flex_dashboard:
orientation: columns
vertical_layout: fill
---
```{r setup, include=FALSE}
library(flexdashboard)
library(magrittr)
library(shiny)
library(ggplot2)
library(dplyr)
```
Column {data-width=650}
-----------------------------------------------------------------------
```{r global, include=FALSE}
set.seed(1234)
df <- data.frame(
var1 = sample(letters, 50, replace = T),
var2 = runif(50),
var3 = sample(LETTERS[1:5], 50, replace = T)
)
df_tmp <- reactive({
var_user <- "A" #session$user
df %>% filter(var3 == var_user) %>% return()
})
```
### Filter
```{r}
radioButtons(label = h4("Choose data"),
inputId = "var1_filter",
choices = df$var1 %>% unique,
selected = df$var1 %>% unique %>% .[1])
```
Column {data-width=350}
-----------------------------------------------------------------------
### Chart B
```{r}
renderPlot({
df_tmp() %>% ggplot(aes(var2)) + geom_density()
})
```
See also:
Create an input variable that is dependent on another input variable in flexdashboard shiny widget